A Conversation With AI

In January 2023, this column labelled 2023 as The Year Of Artificial Intelligence. Since then, the column has explored the application of AI in healthcare and life sciences (AI: Rx For Medical Professionals), the race to create and acquire talent that can leverage AI (Rise Of Generative AI And Demand For Talent To Leverage It), and what the coming generation brought up on AI will be like (GPT Natives: What Will They Be Like?). Generative AI has been capturing everyone’s imagination. The latest research from McKinsey suggests that generative AI can add as much as USD 4.4 trillion annually to just 63 use cases they analysed (for comparison, India’s economy was USD 3.53 trillion in 2022, according to the International Monetary Fund). 

Goldman Sachs says generative AI could automate two-thirds of all jobs and raise global GDP by 7 per cent over the next decade. What the world is becoming is very different from what it once was. How stark is the difference going to be?

To explore the question, let’s roll back a few decades. Let’s see what happened about 80 years ago, in the 1940s. That was the decade in which jet engines arrived, making it possible to circumnavigate the globe in a day. The atomic bomb was unleashed, changing geopolitics. The Z3 was unveiled the first wholly automatic and programmable digital computer with vacuum tubes and switching. 

Though the Z3 didn’t make it past World War II, look where computers are today! DDT, synthetic rubber, aerosol cans, cake mixes, Velcro, the jukebox, the microwave, and television arrived in the 1940s to bring about dramatic change. “The pace of scientific discovery in the 1940s was staggering,” observes Encyclopaedia.com. All, but the jukebox, have survived. The 2020s, the era we live in, aren’t proving too bad either. The James Webb Space Telescope, limb regeneration, heat-tolerant crops, alternative energy sources, and quantum sensors are changing the world in unimaginable ways.

But generative AI has romanced us, momentarily putting all other advances in the shade. Most adults will have their fathers, born around 1940, using smartphones and the internet. But the gulf between them (grandparents) and their grandchildren –who are using generative AI—will be immense as this conversation (imagined—to explore the idea of change) between a grandfather and a teenage grandson shows:

Grandfather: Your sister caught me talking to myself today. Again. Sometimes, it is pretty embarrassing.

Grandson: Hahaha! You do that all the time. It isn’t too difficult to catch you talking to yourself.

Grandfather: Your sister doesn’t know this, but she was talking to herself when she was just three. As a little child, you were doing that too. Most children talk to themselves. It is quite normal—some stop as they grow older. But talking to oneself helps with reasoning, planning, consciously observing, and problem-solving. Research proves this too.

Grandson: If talking to oneself is good, let me create an avatar with your face and voice, plug it into an AI engine, and see whether you like that.

Grandfather: You mean, like, virtually chatting with myself? That used to only be a figure of speech once…

Grandson: Let me set this up for you. We’ll begin by cloning your voice with a few seconds of an audio recording. Just keep talking. (Grandson switches on the recorder on his mobile phone).

Grandfather: (Continues talking) I’ve seen a few sci-fi movies with AI-type characters. I loved Arthur C Clarke’s 2001: A Space Odyssey, but HAL 9000, the AI in the movie, became quite psychotic and murderous. RoboCop was funny. That recent one, Ex-Machina, was disturbing, even a bit terrifying. Blade Runner 2049 was unusual. I mean, Ridley Scott as a director, is the king of neo-noir, but exploring one AI’s relationship with another is stretching it a bit. Do you have any favourites?

Grandson: Not really. But in the next 10 minutes, you are going to be my favourite generative AI character.

Grandfather: (Watching grandson working on the computer) What are you doing now? Grandson: I recorded some 65 seconds of your voice. Now I am creating an online bot and calling it Senior Shailesh 2…

Grandfather: Let me guess, that will be a digital version of me with a mechanical head and comic-book eye movements. Like Shailesh in a Disney production?

Grandson: Give me five minutes. Meanwhile, keep talking to yourself. You have a proven talent for it.

Grandfather: Hahaha! If you were not my grandson, I’d have…

Grandson: Tada! Shailesh 2 is ready. It has been trained using your voice sample. We now have your clone! A few more seconds and we’ll give it your personality… (Grandson enters a prompt, “Grandfather of two, born in 1939, loves George Orwell’s 1984 and still dances to Besame Mucho. Used to work as a telegraph operator in the postal department. Morse code is his thing.”). There we go…say hello to your clone.

Grandfather: (to Shailesh 2, the newly minted avatar) Hello! I am missing Morse code jokes. Tell me one, and we can be friends.

Shailesh 2: (Using Grandfather’s voice) Why did the Morse code message go on a diet?

Grandfather: (Looking a bit puzzled) God knows. You tell me.

Shailesh 2: Because it wanted to be a "light" message!

Grandfather: (Looking amazed) That was a PJ, but it feels bizarre to talk to myself. Just this morning, when I was talking to myself, it felt normal. Now, it feels creepy!

The point of this fictitious but entirely plausible scenario is this: What once felt normal will be redefined. Situations and setups that once were unimaginable will be enabled in seconds—by simple WYSIWYG tools anyone can use to rewrite reality.